Research paper
The Holocene 20(4) 497–516 Quantitative analysis of climate versus © The Author(s) 2010 Reprints and permission: sagepub.co.uk/journalsPermissions.nav human impact on sediment yield since DOI: 10.1177/0959683609355181 the Lateglacial: The Sarliève palaeolake http://hol.sagepub.com catchment (France)
Jean-Jacques Macaire,1 Agathe Fourmont,1 Jacqueline Argant,2 Jean-Gabriel Bréhéret,1 Florent Hinschberger1 and Frédéric Trément3
Abstract Minimum rates of solid (SSY) and dissolved (DSY) sediment yield (SY) were evaluated in t/km2 per yr from sediments stored in the Sarliève palaeolake (French Massif Central) for seven phases of the Lateglacial and Holocene up to the seventeenth century. The catchment (29 km2), mainly formed of limestones and marls, is located in an area rich in archaeological sites in the Massif Central. The respective impacts of human activities and climate on SY were compared by quantification of human settlements through archaeological survey and palynological data. During the Lateglacial and early Holocene up to about 7500 yr cal. BP, variations in SSY and DSY rates were mainly related to climate change with higher rates during colder periods (Younger Dryas and Preboreal) and lower rates during warmer periods (Bölling-Alleröd and Boreal). However, CF1 tephra fallout induced a sharp increase in SY during the Alleröd. During the middle and late Holocene after 7500 yr cal. BP, SSY and DSY greatly increased (by factors of 6.5 and 2.8, respectively), particularly during the Final Neolithic at about 5300 yr cal. BP when the climate became cooler and more humid. After this date, at least 75% of the SSY increase and more than 90% of the DSY increase resulted from human activities, but SSY rates showed little variation during Protohistoric and Historic Times up to the seventeenth century. SSY and DSY rates and DSY/SSY ratio indicate that catchment soils began to form during the Lateglacial and Preboreal, thickened considerably during the Boreal and Atlantic, finally thinning (rejuvenation) mainly as the result of human-induced erosion during the sub-Boreal and sub-Atlantic. Increased mechanical erosion during the late Holocene also induced an increase in chemical erosion.
Keywords archaeological survey, climate-change impact, French Massif Central, Holocene, land-use impact, Lateglacial, palaeolake, soil erosion, solid and dissolved sediment yield
Introduction Current earth surface sediment yield, mainly assessed from solid of lake sediments is often difficult to differentiate and quantify. In and dissolved fluxes in rivers, is highly dependent on human a catchment of given physiography, the ratio of dissolved versus activities whose impact is difficult to quantify compared with that solid yield depends considerably on runoff and thus on vegetation of natural factors, particularly climate (e.g. Judson and Ritter, related to climate and recently to human activities (Walling and 1964; Douglas, 1967; Milliman and Syvitski, 1992; Ludwig and Webb, 1986). Probst, 1998). There are three main questions regarding human An overall increasing sediment yield over the last few thou- impact on sediment yield (Hooke, 2000; Ruddiman, 2003; sand years, with some intermittent periods of lower yields, is Wilkinson, 2005). (1) When did it start? (2) How did it evolve? (3) generally interpreted as the result of human activities (e.g. What was its intensity? These questions can be investigated by Zolitschka, 1998; Hooke, 2000; Edwards and Whittington, 2001). assessing pre-human and syn-human sediment yields, through This interpretation is based on various proxies of anthropogenic studying sedimentary archives stored in favourable basins, par- impact. The most usual human-activity proxy is the palynological ticularly lakes (e.g. Campy et al., 1994; Macaire et al., 1997; Dearing and Jones, 2003). Many studies have shown a trend of increasing sediment yield over several thousand years in popu- 1 Université François-Rabelais de Tours, CNRS/INSU, Université lated areas (e.g. Zolitschka, 1998; Bichet et al., 1999; Dearing and d’Orléans, UMR-CNRS 6113 ISTO, France 2Université Claude Bernard Lyon I, UMR 6636, France Jones, 2003). These studies mainly concern solid sediment yield 3Université de Clermont-Ferrand, EA 1001, France deduced from quantification of detrital sediment stores. Dissolved yield has been more rarely quantified (Einsele and Hinderer, 1998; Received 12 April 2009; revised manuscript accepted 24 October 2009 Gay and Macaire, 1999), although chemical erosion has been found to be of major importance in some lithological or biocli- Corresponding author: Jean-Jacques Macaire, Université François-Rabelais de Tours, CNRS/ matic environments (Meybeck, 1987; Berner and Berner, 1987). INSU, Université d’Orléans, UMR 6113 ISTO, Faculté des Sciences et There is very little information about chemical sediment flux as Techniques, Avenue Monge, 37200 Tours, France compared to the detrital fraction because the precipitated fraction Email: [email protected] 498 The Holocene 20(4)
Surficial formations A Paris B 400 Puy Sarliève paleolake Loire river lake outlet d'Anzelle sediments 528 m Puy tière riv. de Bane Ar 542 m Alluvial terrace I Basalt-rich 400 slope sediments Carbonate-rich A71 slope sediments Sarliève 400 paleolake Bed-rock formations Clermont- A72 Ferrand 345 m Puy de Thiers Miocene basalts Dôme 500 1465m Miocene detrital sediments 600 II Puy de Oligocene volcano- Gergovie sedimentary formations Sancy Issoire Plateau 1885m 745 m Oligocene limestones and marls 700 N I, II: cross-sections 15 km 600 (see Fig. 2) on riv.
Auz 400 Elevation (metres) Sarliève paleolake catchment 1 km
500
Figure 1. Location (A) and main physiographical characteristics (B) of the Sarliève catchment record which provides indirect information (deforestation, cereal INTCAL04 data set. The ‘Marais de Sarliève’ is an ancient lake cultivation, pasture, etc.) from vegetation. Evidence of early (surface area: 5 km2 maximum, altitude: approximately 345 m), human presence is generally provided directly by artefacts, building formed a little before 13 700 yr cal. BP (Fourmont et al., 2006). It structures, etc., but data are often very sparse and were not deter- is located in the Auvergne region, 10 km southwest of Clermont- mined in a systematic manner, and so make it impossible to quan- Ferrand in the Limagne rift (x = 663 780 m; y = 2 082 180 m in tify the density of human occupation and compare it with sediment the Lambert Conformal Conic system; Figure 1A). The origin of yields at different periods. the lacustrine depression is probably tectonic (Ballut, 2000; The aim of this work was thus: (1) to quantify sediment yield Trément et al., 2007b; Fourmont et al., 2009); it can only be (SY) from lacustrine sediment stores for different periods before drained by evaporation or via the outlet located at the northeastern and during human impact since the Lateglacial, distinguishing end of the catchment (Figure 1B). The depression filled with rates of solid sediment yield (SSY) from dissolved sediment yield sediments until it was drained by humans during the seventeenth (DSY), (2) to compare these rates with quantitative data of human century (Fournier, 1996); it is bordered on the north and east by settlement (age, density and location of archaeological sites) from alluvial terraces. The catchment (total surface area: 29 km2) is detailed and systematic archaeological surveys in the lake catch- mainly composed of Oligocene marls and limestones, partly dolo- ment; intensity and type of human activities were evaluated from mitic and sometimes gypsum-rich, with small outcrops of basaltic analysis of pollen grains in lacustrine sediments, and (3) finally to rocks on the surrounding hills up to 745 m (Jeambrun et al., 1973). evaluate the land surface lowering by catchment-soil erosion Bed-rock is generally covered with superficial formations derived which contributed to the sediment yield. from Oligocene carbonated rock (Figure 1B). This area is cur- The studied area is a small lacustrine catchment in the centre of rently characterized by a strong continental climate: mean annual France. Lacustrine deposits, stored since the Lateglacial, have rainfall is less than 600 mm and mean annual temperature is been the subject of detailed sedimentological, palaeohydrological approximately 11°C (min 4°C, max 21°C) (Kessler and and geophysical studies (Bréhéret et al., 2003, 2008; Fourmont, Chambraud, 1986). 2005; Fourmont et al., 2006, 2009; Hinschberger et al., 2006). This Lacustrine sediments, studied through deep pits and core drill- catchment, located in a major site of French history (‘oppidum de ings, consist of delta and basin deposits (Bréhéret et al., 2003, Gergovie’), contains considerable evidence of human settlement 2008; Fourmont et al., 2006, 2009). Delta deposits, 4–5 m thick since the Neolithic and has been the subject of systematic archaeo- (D1 to D7, Figure 2A), are located in the median part of the logical surveys (Trément et al., 2005, 2006, 2007a, b). depression and comprise several units. From base to top these are: D1, beige to greenish carbonated clayey-silts (CS); T1, black pyroclastic sands; D2, several metres-thick deltaic sands com- The Sarliève palaeolake and posed of alternating layers of dark tephric sand and light-coloured its catchment: geological more carbonated sandy-silt; D3, sandy colluvium topped by a setting, sedimentological and palaeosoil; D4, ochre sands; D5, homogeneous greenish calcitic palaeohydrological evolution CS including a thin, pale pink tephra T2; D6, charcoal-rich black calcitic CS; D7, greenish-grey to brown calcitic CS. Basin depos- Here, and in the following sections, all ages are given in calibrated its, 5–6 m thick, observed in the north and south lobes of the years BP. Calibration of radiocarbon ages is based on the depression (B1 to B6, Figure 2A, B), are made up of clayey-silty Macaire et al. 499
A - Cross-section I N-NW Basin area Delta area Basin area S-SE Elevation (m) D6 S23 D7D8 B5 B6 S2 SP4 SP3 SP1 S14 S9 S27 S15 344 C D4-T2-D5 D3 342 B4 P D2 340 B3 D1 B2 B1 400 m T1 338
F - Ca. 11.5 kyrs cal BP - ca. 9.8 kyrs cal BP Outlet Northern Delta Southern B - Cross-section II m basin basin 344 SW NE D3 soil formation 350 Fluvial terrace 342 S3 B1 340 348 S11 C S4 Basin area 338 S5 346 C S6 B6 S7 B5 S8 S26 S9 G - Ca. 9.8 kyrs cal BP - ca 7.5 kyrs cal BP 344 S10 S1 Outlet Northern Delta Southern m basin basin 342 B4 344 P soil formation 340 B3 342 B2 B2 B1 400 m 340 338 338
C - Late Glacial up to 13.7 kyrs cal BP H - Ca 7.5 kyrs cal BP - ca 5.3 kyrs cal BP Outlet Northern Delta Southern Outlet Northern Delta Southern m basin basin m basin basin 344 344 soil formation
342 D1 342 B3
340 340
338 338
D - At ca. 13.7 kyrs cal BP I - Ca 5.3 kyrs cal BP - ca 2.3 kyrs cal BP Northern Southern m Outlet Delta Outlet Northern Delta Southern basin basin m basin basin 344 344 D4-T2-D5
342 T1 342 B4
340 ? 340 ? 338 338
E - Ca. 13.7 kyrs cal BP - ca. 11.5 kyrs cal BP J - Ca 1.8 kyr cal BP - ca 0.3 kyr cal BP Northern Delta Southern Outlet Northern Delta Southern m Outlet m basin basin basin D6 D8 D7 B6 basin 344 344 B5 342 D2 342
340 B1 340
338 338
maximum water level probable usual water level
Figure 2. Cross-sections through sediments (A, B) and pattern of Sarliève palaeolake evolution since the Lateglacial up to the seventeenth century (C to J) (from Fourmont et al., 2009). Location of I and II cross-sections, see Figures 1 and 4B. S2, S9, core drilling; SP3, SP4, pit section; D1 to D7, delta deposits (T1 and T2, tephras); B1 to B6, distal basin deposits; P, proximal basin deposits; C, colluvium 500 The Holocene 20(4)
sediments, generally carbonate-rich, showing from base to top: Age Sediment Lacustrine sediment oxides B1 unit, quartz-rich compact yellowish to greenish clayey-silts (in yrs cal BP) units (as % of the bulk) (CS), characterized by the presence of zeolites; B2, grey to blue 0 20 40 60 80 100 0 clayey-silts mainly composed of authigenic dolomite; B3, alterna- tion of dark calcitic CS (cm-thick layers) with bundles of dolo- B6 mitic and aragonitic laminae; B4, homogeneous greenish calcitic B5 CS; B5, charcoal-rich black calcitic CS; B6, greenish-grey to 2000 hiatus brown calcitic CS. The lower part of proximal basin deposits con- tains a dark sandy layer incorporating reworked pyroclastic parti- cles (similar to T1 and D2). The following palaeohydrological and sedimentological scheme 4000 B4 of the lake’s evolution (Fourmont et al., 2009) has been drawn up Detrital from distribution and comparison of sedimentary facies, 14C datings, oxides tephra ages and detailed mineralogical and geochemical analysis. During the Lateglacial up to c. 13 700 yr cal. BP (mainly 6000 Bölling), D1 unit sediments, containing c. 30% travertine-like B3 precipitated material, were deposited in the delta area; no evi- dence of coeval sediments was found in basins. Water level was 8000 CaO MgO low, with a maximum depth of c. 5 m (Figure 2C). At c. 13 700 prec prec yr cal. BP, at around the beginning of the Alleröd, direct CF1 B2 tephra (Vernet and Raynal, 2000) was deposited (T1 unit, Figure 2D). While it probably covered the whole catchment area, CF1 10000 tephra has only been observed in delta sediments where it was SiO2 + buried under the D2 unit deposited in a fan-type delta during the B1 Al2O3 period 13 700 yr cal. BP–11 500 yr cal. BP (Alleröd and Younger prec Dryas) (Figure 2E). The D2 unit, exclusively detrital in origin, 12000 includes layers of tephra reworked from catchment slopes, sug- gesting a rise in water level during flood event(s) to more than c. 342 m, which probably led to a short-lived opening of the Figure 3. Evolution of contents in detrital and precipitated fractions in Sarliève palaeolake basin sediments (core S17). CaOprec, MgOprec lacustrine depression. However, during the same period, from and SiO + Al O prec, precipitated CaO, MgO and SiO + Al O in 13 700 yr cal. BP up to c. 9800 yr cal. BP, authigenic zeolites 2 2 3 2 2 3 percents of lacustrine bulk sediments (from Fourmont et al., 2009). and quartz formed in the distal basin B1 unit, partly because of SiO + Al O were precipitated from CF1 tephra weathering alteration of CF1 tephra, and indicating endorheic conditions 2 2 3 with a generally low water level and reduced detrital input (c. 60% of B1 sediments) (Figure 3). detrital input during periods with low water level and salty water, After c. 11 500 yr cal. BP (early Holocene), colluvium (D3 ending with the input of fresh water. This trend could be related to unit) was deposited on the emerged delta area (Figure 2F). Soil repeated time-limited fluctuations of the water level which was formed at the top of D3, probably during the early and middle generally below the top of D3 palaeosoil, i.e. c. 342/343 m. The Holocene until the end of the Atlantic period (Figure 2G, H). duration of one BL + HCS deposition sequence has been evaluated During this period (9800 yr cal. BP until 7500 yr cal. BP) at 150 yr on average in B3 and 470 yr at the maximum between authigenic dolomite-rich sediments (B2 unit) formed in the 7100 ± 150 yr cal. BP and 7260 ± 160 yr cal. BP. However, the basins, indicating perennial endorheic conditions. The bio- gradual thickening of detrital HCS facies towards the top of B3 chemical dolomite content increased in the B2 sediments from indicates the increasing input of fresh water and the possible tem- 30% (bottom) to 60% (top), indicating maximum restriction at porary opening of the system. around 8000 yr cal. BP (Figure 3). The water level did not The sediment geometry shows that the water level was high reach the top of D3, and the water was probably quite shallow during the late Holocene from 5.3 ka cal. BP: B4 to B6 units accu- (less than 1–2 m deep). mulated when the lake extended, covering the delta area and border During the middle Holocene, from c. 7500 yr cal. BP until c. deposits (Figure 2I, J). Water-level altitude must have been at least 5300 yr cal. BP (Atlantic period), when the delta area was never 343 m at the beginning of the sub-Boreal, and 344/345 m in the flooded (Figure 2H), unit B3 was deposited in the basins. seventeenth century. Apart from the D4 sandy transgressive facies Interbedding of laminated biochemical sediments (BL facies) with and local D8 sandy-gravelly deltaic facies, the upper sediments of detrital sediments (homogeneous clayey-silts: HCS facies) indi- the whole delta (D5 to D7 units) and basin (B4 to B6 units) have a cates variations in palaeohydrology and related depositional con- homogeneous composition. They contain more than 80% detrital ditions. Bréhéret et al. (2008) identify three phases of BL–HCS materials, and biochemical calcite is always < 20% (Fourmont et al., facies sedimentation: (1) dolomite authigenesis in BL under 2009; Figure 3). The waters were probably continuously fresh, restricted conditions similar to B2; (2) aragonite genesis in BL, implying a perennial body of water in the lake, total opening of the indicative of Ca2+-rich freshwater inputs when facies are com- depression and continuous water supply from the catchment. posed of 20% to 60% biochemical phases (Figure 3) and deposited However, archaeological data (Gallo-Roman settlements, cemetery while lake waters were salty; and (3) HCS facies, where bio- and drainage networks in the palaeolake basins) show that the lake chemical phases constituted less than 10%, indicating increased dried up between the third century bc and second century ad Macaire et al. 501
because of human activities (Trément et al., 2007b). Moreover, and D7-B6 (greenish-grey to brown calcitic CS) units could not ‘catastrophic’ flood(s) occurred between the fourth and seventeenth be distinguished (Figure 2). centuries ad, causing sandy-gravelly fan delta deposition (D8 unit), The apparent volumetric mass (dry weight divided by the wet and resulting in the water level rising above 346/347 m. Finally, the volume, in t/m3) of the basin-unit sediments was calculated by lake was drained during the seventeenth century (Fournier, 1996), weighing sediments after given volumes of sediments were dried and since then the water has run in artificial channels. at 40°C for 48 h. We decided that the apparent volumetric mass of lower sandy delta units, which could not be measured, was 1.5 t/m3, based on data of Macaire et al. (2006). The sediment mass for Methods each unit was calculated by multiplying the unit’s volume with the experimentally derived volumetric mass. From these data we cal- Evaluation of catchment physiographical culated the mass of material (detrital and precipitated) forming characteristics and calculation of sediment each lithological unit, using the data of Fourmont et al. (2009). yield rates For B3 unit, composed of alternating precipitated material-rich Some parameters of relief and lithology were analysed using GIS. laminae (BL) and detrital material-rich layers (HCS), the whole Slope gradient and true surface area of the catchment were calcu- sediment mass was calculated from the volumetric mass of BL and lated from DEM using Arcview 8.2 software and digitizing eleva- HCS facies weighted by their relative thickness in the unit. The tion contour lines of Institut Géographique National (IGN) 2531E evolution of the mass of precipitated material compared to detrital and 2532E 1/25 000 maps. Surface areas of main types of rocks material from bottom to top in B3 was deduced from evolution of were assessed by digitizing XXV-31 ‘Clermont-Ferrand’ 1/50 000 the thickness of BL compared with HCS facies. geological map (Jeambrun et al., 1973). Age and period of formation of each lithological unit were Sediment yielded by catchment was calculated in t/km2 per yr calculated based on the chronological model of Fourmont et al. from stored lacustrine sediment mass. Volumes of main sedimen- (2009), that is based on 18 calibrated 14C ages of seeds, charcoal tary units (delta sediments: units D1 to D8, and T1 tephra; basin and bulk sediment organic matter, performed by the CDR (Centre sediments: units B1 to B6) were calculated from their thickness de Datation par le Radiocarbone, Lyon 1 University), completed and extent using different methods. Facies and thickness of sedi- with information from pollen zones, tephrochronology and ments were observed directly (1) from 13 pit sections of 8 m × 8 m, archaeological data. The base of D1 was arbitrarily chosen as the carried out up to a depth of 3.3 m by the Institut National de beginning of the Bölling (15 000 yr cal. BP). Deposition of D3 Recherches Archéologiques Préventives (INRAP) for an archaeo- colluvium was attributed to the Preboreal (duration: 1100 yr), and logical survey of the construction site of the ‘Grande Halle palaeosoil formation on this colluvium to the Boreal and Atlantic d’Auvergne’, located near by the delta area, and (2) from 31 core (duration: about 5100 yr). drillings retrieved with a percussion sampler (Eijkelkamp) by the Sediment source area for calculation of the sediment yield was geological laboratory of Tours (UMR CNRS 6113 ISTO) in north- determined for each unit as the difference between the total catch- ern and southern distal basins (Figure 4B). Sediment thickness and ment area and the unit surface area in the lacustrine area. extent were determined using 56 geotechnical tests (static pene- Sediment yield rates were estimated in t/km2 per yr by dividing trometry using 100KN penetrometer) carried out by Fondasol- the each unit’s sediment mass by the surface of sediment source Auvergne: penetrometric data mainly identified hard substratum and the time period of formation of the unit. Total sediment yield (resistance to penetration R > 6 Mpa), and with regard to lacus- (TSY) was calculated, as well as solid (SSY) and dissolved trine sediments, clayey silty (B1 to B6, D1, D4 to D7 units: R < 2 (DSY) sediment yields. It was possible to distinguish between MPa), sandy-silty (T1, D2 and D3 units: 2 < R < 6 MPa) and detrital and precipitated material in sediments thanks to geo- gravelly-sandy (D8 unit: R > 6 Mpa) bodies, according to data of chemical, XRD, SEM and EDS analyses. Mineral study suggests Antoine and Fabre (1980). that, except in B1, silicates (quartz, feldspars, zeolites and clay In the northern and southern basins about 50 km of electro- minerals) are mainly detrital, and carbonates (calcite, dolomite magnetic profiling was performed using an average recording and aragonite) are both detrital and in situ precipitated (biochem- distance of 7 m along the profiles, a lateral spacing of 20 m to ical). For each unit, the mass of precipitated MgO and CaO car- 100 m, an EM 31 instrument, and the slingram method (Mc Neil, bonates have been quantified (details in Fourmont et al., 2009), 1980) (Figure 4C). Electrical conductivity of surficial deposits up (1) from MgO and precipitated dolomite contents in sediments to about 6 m depth was calibrated with lithological facies determined from geochemical analyse and SEM observation, observed from core drillings. Sixteen electrical soundings were respectively, (2) by comparing CaO contents in sediment and also performed. Based on the geophysical survey, the total thick- catchment soils for each period. From this investigation, the ness of lacustrine sediments in basins and the thickness and uncertainty in determination of SSY and DSY values can be extent of sediment units mainly composed of precipitated (B1 evaluated to ± 20%. to B3) and detrital material (B4 to B6) were determined (Hinschberger et al., 2006). Spatial analysis of these data (eleva- tion of bottom and top, and 3D extent of units) were performed Pollen study using SIG (Arcview 8.2; Fourmont, 2005). Interpolation by krig- Pollen analysis (pollens data used for the chronological model of ing was used to delineate the shape of lacustrine sediment units, Fourmont et al., 2009) was carried out on samples collected by based on which sediment volumes were calculated. It was not increments of 3 cm in P core performed with Russian core sampler possible to distinguish the limit between T1 (tephra) and D2 in the northern basin (Figure 4B). After eliminating carbonate and (tephra-rich delta sediments) units, as their grain sizes are similar. siliceous components of sediments, pollen grains were concentrated Likewise, clayey-silty D6-B5 (charcoal-rich black calcitic CS) with ‘Thoulet’ solution (d = 2) (Goeury and de Beaulieu, 1979). 502 The Holocene 20(4)
N A 0km 1
archaeological A75 highw survey area
ay Puy de tière Bane L’Ar
catchment limit
Sarliève paleolake rÈment T O F. DA Elevation 700-750 Plateau de 650-700 Gergovie 600-650 550-600 500-550 450-500 L’Auzon 400-450 350-400 < 350 A75 A75 non-destructive core drilling electrical soundings B (S) (UMR CNRS 6113) C electromagnetic profiles P core for palynology (EM31) Gachon core (1963) D212 D212 S23 limit of electromagnetic pit section for archaeological prospecting survey survey (INRAP) P static penetrometry survey Conductivity (mS/m): by Fondasol < 100 cross-sections (see Fig.2A, B) 100 - 150 “Grande rase” 350 m 150 - 200 350 m catchment limits 370 m 200 - 250 370 m elevation curves (m) 370 m 370 m D137 D137
350 m 350 m
S3 390 m 390 m
N N
S11 0 km 1 S15 0 km 1
Figure 4. Location of surveys. (A) Archaeological survey area in whole catchment; (B) location of core drillings, pit section for archaeological survey and static penetrometry survey in palaeolake area; (C) geophysical survey in palaeolake area
A minimum of 300 pollen grains were identified and counted in Historical study and archaeological survey each sample. To make it easier to interpret pollen, taxa were sepa- rated into two classes: those characteristic of wetland (area in The history of human settlement was investigated through exhaus- direct contact with the Sarlieve palaeolake), and those characteris- tive bibliographical study of the nine districts in the catchment. tic of the catchment. The abundance of pollen grains was calcu- Five systematic archaeological field surveys, with intervals of 10 m lated as a percentage of each class and analysed separately. and sampling of the whole archaeological material on each site, Chenopodiaceae, which are not strictly related to either class, and were performed between 2001 and 2003, in accessible areas with Tertiary taxa were calculated as a percentage of the whole pollen no buildings, representing 40% of the whole catchment area grains. Data on non-pollinic microfossils studied on P core were (Trément et al., 2007b; Figure 4A). Four spatial archaeological taken from Argant and Lopez Saez (2004). units were identified for different periods: (1) site, which has a Macaire et al. 503
concentration of artefacts providing evidence of human activity at A Slope surface area (as %) Cumulated slope surface area (as %) a precise place, (2) probable site, which has insufficient evidence, 50 100 (3) isolated artefact, for a notable artefact, and (4) off-site pottery, 40 80 representing a ‘background noise’ of scattered and often sub- 30 60 rounded abundant artfacts resulting from farming-related soil 20 40 enrichment. Statistical analyses of the sites interpreted as settle- ments were carried out for each period: number, distribution by 10 20 elevation and total surface area. In addition, the 13 pit sections 0 0 0-2 3-4 5-7 8-10 11-13 14-16 17-19 20-24 25-30 31-49 carried out by INRAP before the construction of the Grande Halle Slope-gradient classes (in°) d’Auvergne (Vernet et al., 2005) were studied; they cover a sur- face area of 90 ha in the palaeolake delta zone at its eastern and Paleolake Catchment (2) 2 B (1) western borders where there is a large site (7500 m ) (Figure 4B). Surficial formations Bed-rock formations
65 67.7 Results 60 55 Catchment physiography 50 Mean slope gradient(in°) Catchment slope gradients vary from about 0° in the Sarliève 25 25 depression to 49° on the eastern side of the ‘Plateau de Gergovie’ (Figure 5A). Slope gradients are lower than 2° on 40% of the catch- 20 20 ment area and above 8° on about 20%. Percentages of main rock- 15 17.4 15 14.8 outcrop surfaces are shown in Figure 5B: carbonated rocks cover 10 10 about 75% of the catchment in low to medium slope gradient areas. Outcrop surfaces area (as %) 5 6.8 5 5.0 3.1 0.2 0 2.4 0 Sediment yield (SY) Rates of total (TSY), solid (SSY) and dissolved (DSY) sediment
Sarliève Basalt-rich Carbonate-rich Miocene yields, and data for their calculation are shown in Figure 6A,B,C and paleolake sedimentsFluvial sedimentsslope sediments slope sedimentsMiocene basalts Oligocene volcano- detrital sediments sedimentaryOligocene limestones in Tables 1 and 2, respectively. During the Lateglacial and Holocene, formations and marls SSY and DSY varied from 6 to 203 t/km2 per yr and from 4 to 44 Figure 5. Physiographical data of Sarliève catchment. (A) Catchment t/km2 per yr, respectively. Nine sediment yield phases (SYp1 to SYp9) slope gradients. (B) Lithology and mean slope gradients by lithological 2 were identified. During the Bölling and at the beginning of the Alleröd area. (1) lithology as per cent of whole catchment area (29 km ); (2) lithology as % of sediment-yielding catchment area (24 km2) (SYp1), SSY (19 t/km2 per yr) and DSY (8 t/km2 per yr) rates were low. During the second part of the Alleröd and the younger Dryas (SYp2), SSY (65 t/km2 per yr) and DSY (31 t/km2 per yr) increased Palynology significantly. During the Preboreal and at the beginning of the Bölling Palynological data obtained from study of the P core concern B3 (SYp3), SSY decreased slightly (52 to 46 t/km2 per yr), while DSY did to B6 lithological units (Atlantic to late sub-Atlantic). The taxa, not vary. The lowest SY rates appeared during the Bölling and at the described in detail in Argant and Lopez Saez (2004), are presented beginning of the Atlantic (SYp4): SSY decreased from 11 to 6 t/km2 in this paper in two environments: catchment and wetland (Figure 7). per yr, while DSY increased from 4 to 9 t/km2 per yr. During the This made it possible to investigate the relationship between SY Atlantic and at the beginning of the sub-Boreal (SYp5), there was rates and plant cover during anthropogenic impact periods (Ep5 to considerable variation between SSY and DSY rates (24 to 55 t/km2 per Ep8). These elements further Gachon’s data (1963) on vegetation yr and 37 to 6 t/km2 per yr, respectively): SSY and DSY increased in the Sarliève area for the whole Holocene and are in line with alternately during this period, with a general pattern of SSY increase Prat’s data (2006) for the sub-Boreal and sub-Atlantic. and DSY decrease. During the sub-Boreal and at the beginning of the Catchment plant cover (Figure 7A) was dominated by forest sub-Atlantic (SYp6), SSY increased considerably (161 t/km2 per yr), (50–85% of arboreal pollen) up to the end of the sub-Boreal. and DSY rose to 40 t/km2 per yr. The SY rate for the period between During the Atlantic, Quercus and Corylus dominated (pollinic 2300 and 1800 yr cal. BP (SYp7) could not be calculated as there were zones a, b and c) with Ulmus and Tilia; during periods of reduced no stored sediments due to lake-water drainage at this time. SY rates arboreal pollen grains, Corylus decreased more than Quercus. After were highest during the sub-Atlantic from 1800 to 300 yr cal. BP that, Fagus and Abies appeared (zone d) and developed during the (SYp8), up to 203 t/km2 per yr between 1600 and 300 yr cal. BP sub-Boreal (zones e, f, g and h), mainly to the detriment of Corylus. (SYp8b) for SSY, and 44 t/km2 per yr for DSY. SY could not be evalu- At the beginning of the sub-Atlantic there was a sharp decline of ated for the period after 300 yr cal. BP because of the lake drying up. forest taxa (always lower than 40%: zones i and j), dominated by SSY was generally greater than DSY, except at the end of SYp4 Pinus. Poaceae were always present (5 to 20% of pollen grains in and the beginning of SYp5. The DSY/SSY ratio (Figure 6D) the catchment), but dominated (up to about 50%) only in zone i, at remained close to 0.5 with a slight increase (0.42 to 0.67) between the beginning of the sub-Atlantic. Ruderals (Plantago, Artemisia, SYp1 and SYp3. This ratio increased sharply to 1.54 during SYp4 Centaurea, Cichorioideae, Asteraceae, Polygonum, Lamiaceae, (beginning of the Atlantic), then decreased with variations to 0.11 Apiaceae), except Chenopodiaceae, were found from the beginning at the end of SYp5. After that, the DSY/SSY ratio did not vary of the Atlantic (zone a); they were temporarily more abundant in significantly, remaining around 0.25 during SYp6 and SYp8. zone c with a marked increase at the sub-Atlantic (up to 45% in 504 The Holocene 20(4)
A B C D Dates (yrsChronozones cal SYBP) phases 0 SYp9
1000 SYp8 247 b ? ? Sa 222 2000 SYp7 a
3000 ? ? 4000 Sb SYp6 201
5000
6000 SYp5 61 ? ? 7000 At
8000 SYp4 15 9000 Bo
10000 77 b SYp3 11000 Pb 83 a
12000 Y D SYp2 96 13000 Al 14000 Böl SYp1 27 15000 0 50 100 150 200 250 250 200 150 100 50 0 50 0 0.5 1 1.5 Total sediment yield Solid sediment yield Dissolved Ratio dissolved/solid sediment yield sediment yield (TSY in t.km−2.yr−1) (SSY in t.km−2.yr−1) (DSY in t.km−2.yr−1)
Figure 6. Rates of sediment yield (SY) during the Lateglacial and Holocene in Sarliève catchment. Chronozones from Boivin et al. (2004). (A) Total sediment yield, (B) solid sediment yield, (C) dissolved sediment yield, and (D) ratio of dissolved/solid sediment yield zone j). Cereals, not observed in pollinic zone a, appeared as traces surface within the Sarliève catchment (Figure 8A). Evidence for in zone b and remained present, with pollen-grain contents tempo- the early Neolithic (Recent Cardial) has been found, although poor, rarily higher during the sub-Boreal in zones e and f and up to about while evidence for the late/final Neolithic, difficult to identify, is 20% of catchment taxa in zone j. sparse and mainly located in high elevation areas. Evidence of the Wetland vegetation (palaeolake edge) was also characterized middle Chassean Neolithic dominates (17S, 3PS and 5IA), mainly by the predominance of arboreal taxa (mainly Alnus, with Fraxinus located on slopes of the Plateau de Gergovie on the west and slopes and Salix) during the Atlantic and sub-Boreal (Figure 7D), with of the Puy d’Anzelle and Puy de Bane on the east of the catchment. greater and more frequent variations in the abundance of arboreal There is particularly abundant evidence of the ancient Bronze taxa during the Atlantic (30–80% arboreal pollen grains domi- Age (14S, 6PS and 1IA), especially on the Plateau de Gergovie nated by Fraxinus: zones a to d) than during the sub-Boreal (55 to and Puy d’Anzelle slopes, and always at an elevation above 360 m 85%, dominated by Alnus: zones e, f and g). Arboreal taxa (Figure 8B). There is some evidence of the middle Bronze Age decreased sharply at the end of this period (5%, bottom of zone i). (5S, 1PS and 1IA) at the same location as evidence of the ancient During the sub-Atlantic, arboreal taxa first increased (up to 50%, Bronze Age and above 370 m. There is abundant evidence of the zone i), then decreased sharply (5–20%, zone j). Areas without final Bronze Age (10S, 3PS, 1IA) and 1 funerary site (FS); they forest were mainly covered by Cyperaceae which were particu- are mainly located on western and eastern catchment slopes, and larly abundant at the end of the sub-Boreal (zone h) and at the the funerary site is the first evidence of human settlement in the beginning of the sub-Atlantic (zone i). Aquatics sometimes appear. Sarliève depression at an elevation of 346 m. Chenopodiaceae (Figure 7C) were always present, but more Evidence of the first Iron Age is very abundant (17S, 7PS), abundant at the beginning of the Atlantic (zone a: up to 15% of mainly on slopes, from the eighth century bc. This evidence is not total pollen grains), temporarily during the sub-Boreal (zones f found on the plateau, but in areas close to the palaeolake. There is and h) and sub-Atlantic (zones i and j). Tertiary taxa (mainly only a little evidence of the ancient and middle La Tène (fifth–third Gymnosperma) were always present (Figure 7B), often forming centuries bc) (7S, 4PS, 1IA and 3FS), mainly found at low elevation 10% of total pollen grains, particularly during periods of catch- (387 m on average). Human pressure increased sharply during the ment arboreal taxa decrease, but they were as high as 90% in final La Tène (second–first centuries bc) (Figures 8C and 9): 24S, pollinic zone e at the beginning of the sub-Boreal. 15PS and 2FS, especially at lower elevations (53% of evidence is found below 350 m), at the edge of the eastern alluvial terrace close to the palaeolake and in the northern outlet area, although the whole Archaeology catchment was settled by humans at that time. Off-site pottery Surface archaeological survey results have been described in detail resulting from farming-related soil enrichment is closely associated by Trément et al. (2007b). They are summarized in Figures 8 and with sites. Land parcels with boundary markings and drainage 9. Evidence for the Neolithic is found in 21 sites (S), 7 probable ditches, dating from the second half of the first century bc, have sites (PS) and 12 isolated artefacts (IA) observed on the land been discovered at the bottom of the lacustrine depression. Macaire et al. 505
The early Roman Empire (first–second centuries ad) is characterized by a sharp increase in signs of settlement (57S, per yr) 2 31PS and 11FS) throughout the catchment (hilltops, slopes, edge (t/km SSY (3) SSY 25 160 0.5 6.1(8.4)10.7 6 46 19 24.6(39.9)55.3 19 178 and bottom of lacustrine depression); mean elevation of sites is low: 390 m. The off-site pottery area is widespread. Sites were either vast agricultural domains (or villae) (Trément, 2007b), or
per yr) buildings belonging to the villae, or small farms. During the late 2 Roman Empire (third–fifth centuries ad) (Figures 8D and 9), only DSY (2) DSY 0 40 0 4.6(6.9)9.2 0 31 8 (t/km 6.1(21.5)36.8 0 44 17 of the former 57 sites remained, mainly located at the edge of the palaeolake (mean elevation: 383 m), in the same place as the main early Roman Empire buildings (villae). During the early Middle Ages, settlements were still few (5S, ) 2 4PS, 2IA and 2FS) and always located at very low elevations (mean: 351 m): slopes and plateaus seem to have been deserted. SY (1) surface SY 23.8 24.9 24.9 26.6 24.9 27.3 25.8 area (km 25.6 24.9 24.5 The off-site pottery area was also less extensive than during the early Roman Empire. Four of the five sites were occupied con-
t) tinuously from the final La Tène and became vast seigniorial 6 domains during the Middle Ages. Archaeological excavations carried out before the construction 0.773 0.038 0.937 0.153 0.819 0.921 3.457 1.014 8.165 Mass (10 14.98
of the ‘Grande Halle d’Auvergne’ (Vernet et al., 2005; Trément et al., 2007b) discovered a funerary site with a tumulus dating ) 3 from the final Bronze Age and extensive settlement between the third century bc and second century ad at the northeastern edge of the palaeolake. During the La Tène period, settlement, comprising Volumetric 1.5 mass (t/m 1.5 1.75 1.5 1 0.95 1.75 0.75 1.5 2 workshops, wells and burial places, was highly structured with enclosures and drainage networks. There is evidence of human ) 3
m activity (wells, parcels with boundary marking) on the northern 6 lacustrine depression bottom at the time of the Roman conquest (Gergovia battle: 52 yr bc) and during the early Roman Empire (buildings, collection of spring water, funerary sites) up to the 0.102 Volume (10 Volume 0.516 8.56 0.025 0.937 0.862 0.526 4.61 0.676 4.083 third century ad. Evidence of fisheries dating back to the Middle Ages and Modern Times have also been found.
90 Discussion > 80 60 70 80 Significance of rates and phases of 100 Detrital components (%) 100 100 (very variable) 70 to 40 (from bottom to top) 40 to 100
sediment yield Calculated TSY, SSY and DSY rates are averages for the whole sediment yield surface area (27.3 to 23.8 km2 according to period, Table 1). In the catchment, spatial distribution of SY varied according to earth surface conditions, particularly slope gradient 10 to 60 0 0 0 0 Precipitated components (%) 20 < (very variable) 30 to 60 (from bottom to top) 40 30
20 (0–49°) and lithology (dominance of basalt or limestone and marl) (Figure 5). Moreover, SY rates are averages of SY phases whose duration depends on datation possibility and during which SY could vary due to catastrophic climatic events when retention rates of particles in the lake might have been low (Brune, 1953). Nine 1100 ? 1100 Deposition duration (yr) 1300 3000 3000 2200 2300 3900 1300 ? 2200 1500 sediment yield phases (SYp1 to SYp9) were identified covering the main chronozones of the Lateglacial and Holocene (Figure 6). Calculated SSY rates (Figure 6B) are minimum rates which can be considered close to true rates because many factors are favourable to high palaeolake-trap efficiency (Heineman, 1984;
1.6–0.3 5.3–2.3 5.3–2.3 7.5–5.3 9.8–7.5 1.8–0.3 Age (kyr cal. BP) 11.5–10.4 ? 11.5–10.4 13.7–9.8 15.0 ?–13.7 13.7–11.5 Einsele and Hinderer, 1998): high lake/catchment area ratio (aver- age of 0.16), lake often closed or shallow when open, lake water rich in Ca2+ and Mg2+ favouring clay flocculation, and inflow mainly due to runoff with no river. During the Lateglacial and basin B4 B3 B2 B1 B5+B6 early and middle Holocene (SYp1 to 5), endorheism is shown in Data for sediment yield calculation, and rates of solid dissolved per lacustrine unit. (1) Sediment yield; (2) (3) basin sediments by abundance of minerals (analcite, dolomite, D7 D2 aragonite) resulting from chemical or biochemical precipitation in + +
Table 1. Table Sediment units delta D8 D6 D5 D4 D3 T1 D1 salty water. Particle retention in the palaeolake was favoured by 506 The Holocene 20(4)
Table 2. Rates of solid and dissolved sediment yield per sediment yield (SY) phase. (1) Dissolved sediment yield; (2) solid sediment yield; (3) total sediment yield
SY phases Ages (kyr cal. BP) Sediment units DSY (1) (t/km2 per yr) SSY (2) (t/km2 per yr) TSY (3) (t/km2 per yr) DSY/SSY ratio
SYp9 0–0.3 no sediments ? ? ? ? SYp8 b 0.3–1.6 D6+D7+D8+B5+B6 44 203 247 0.22 a 1.6–1.8 D6+D7+B5+B6 44 178 222 0.25 SYp7 1.8–2.3 no sediments ? ? ? ? SYp6 2.3–5.3 D4+D5+B4 40 161 201 0.25 SYp5 5.3–5.5 B3 6 55 61 0.11 5.5–7.5 B3 37 24 61 1.54 SYp4 7.5–8.2 B2 9 6 15 1.50 8.2–9.8 B2 4 11 15 0.36 SYp3 b 9.8–10.4 B1 31 46 77 0.67 a 10.4–11.5 D3+B1 31 52 83 0.60 SYp2 11.5–13.7 T1+D2+B1 31 65 96 0.48 SYp1 13.7–15.0 D1 8 19 27 0.42
ones Depth A B C D (in cm) Catchment vegetation Tertiary taxa Chenopodiaceae Wetland vegetation Sediment Localones pollinic 0 units z Chronoz SY phases cereals aquatics
ruderals Cyperaceae B6 j and paludals 8
Poaceae B5 i sub-Atlantic
100 h
g
trees trees 6 B4 f sub-Boreal
200 e
d
300 c 5 B3 Atlanti c
b
a 7750 yrs cal BP 0 50 100 0 50 100 0 30 0 50 100 %
Figure 7. Pollen diagrams from P core (location, see Figure 4B). (A) Catchment vegetation as per cent of total catchment taxa. Trees: Corylus, Quercus, Ulmus, Tilia, Betula, Fagus, Abies, Juniperus, Hedera, Acer, Picea, Pinus. Ruderals (except Chenopodiaceae): Artemisia, Asteraceae, Carduus, Centaurea, Cuscuta, Brassicaceae, Cichorioideae, Convolvulaceae, Caryophyllaceae, Malvaceae, Apiaceae, Plantago, Rumex, Rubiaceae, Fabaceae, Lamiaceae, Urticaceae, Polygonum. (B) Tertiary taxa as per cent of total pollen grains. (C) Chenopodiaceae as per cent of total pollen grains. (D) Wetland vegetation as per cent of total wetland taxa. Trees: Alnus, Salix, Fraxinus, Ligustrum, Viburnum, Myrica, Sambucus, Vitis. Cyperaceae and paludals: Filipendula, Thalictrum, Typhaceae. Aquatics: Potamogeton, Lemna
endorheic conditions (Einsele and Hinderer, 1998) which were the fact that only precipitated low-Mg calcite is found in sediments continuous until the beginning of the Atlantic and then intermit- is the consequence of open conditions – an extensive freshwater tent until the beginning of the sub-Boreal (Bréhéret et al., 2003, lake with fluctuating water level, as shown by non-pollinic micro- 2008; Fourmont et al., 2006). During the late Holocene (SYp6 and 8), fossils (Argant and Lopez-Saez, 2004), and probable frequent Macaire et al. 507
A 0km 1 B 0km 1
Puy de Puy de Bane Bane L'Artière L'Artière
Sarliève Sarliève paleolake paleolake F. TrÈment F. TrÈment Elevation Elevation 700-750 700-750 650-700 650-700 Plateau de Plateau de 600-650 600-650 Gergovie Gergovie catchment limit 550-600 catchment limit 550-600 500-550 500-550 450-500 450-500 L'Auzon 400-450 L'Auzon 400-450 350-400 350-400 < 350 < 350
Ancient Neolithic: Probable site Ancient and Middle Bronze Ages: Site Probable site Isolated artefact Middle Neolithic: Site Probable site Isolated artefact Final Bronze Ages until Ancient La Tène: Site Probable site Isolated artefact Funerary site
C 0km 1 D 0km 1
Puy de Puy de Bane Bane
L'Artière L'Artière
Sarliève paleolake Sarliève paleolake F. TrÈment F. TrÈment Elevation Elevation 700-750 700-750 650-700 650-700 Plateau de 600-650 Plateau de 600-650 Gergovie 550-600 Gergovie catchment limit 550-600 500-550 500-550 catchment limit 450-500 450-500 L'Auzon 400-450 L'Auzon 400-450 350-400 350-400 < 350 < 350 Final La Tène: Late Roman Empire: Agglomeration Site Probable site Funerary site Large villa Small villa Site Probable site Funerary site Roman road Off-site pottery Early Roman Empire: Large villa Large villa with Final La Tène traces Isolated artefact Early Middle Ages: Small villa Small villa with Final La Tène traces Roman road Site Probable site Funerary site Isolated artefact Site Probable site Off-site pottery Sanctuary Funerary site
Figure 8. Location maps of human settlements in Sarliève catchment at different periods. (A) Ancient and middle Neolithic; (B) Bronze Ages up to ancient La Tène; (C) final La Tène and early Roman Empire; (D) late Roman Empire and early Middle Ages outflow and loss of solid matter downstream of the lake during abundant slope formations (74.5% of catchment area: Figure 5B) highstand periods. Although sediment storage increased markedly are mainly due to periglacial processes (rock fall, solifluxion: at this time (Table 1), SY may be slightly underestimated. Very Jeambrun et al., 1973) and thus prior to lake formation. Colluvium 508 The Holocene 20(4)
A short periods (SYp5), then more continuously (SYp6 and 8) when 60 the palaeolake spread out and detritism increased, DSY rates are
50 probably grossly underestimated, although they were high during the sub-Boreal and sub-Atlantic (Figure 6C). 40 The Sarliève catchment is located in medium-sized mountains 30 where slope gradients are low (< 5° for 60% of the catchment area; 20 Figure 5) and carbonated rocks dominate (about 75% of the area; 10 Figure 1B). SSY and DSY rates for the most recent period (178 to 203 and 44 t/km2 per yr, respectively for SYp8) are a little higher 0 than those assessed from current river flux (Meybeck, 1979; e-Iron Milliman and Meade, 1983) and as high as soil erosion rates e-Bronze l-Bronze e-Empirel-Empire l-Neolithic m-Bronze e-La Tène l-La Tène e-Neolithicm-Neolithic m-La Tène (Cerdan et al., 2006) in similar physiographic and bioclimatic e-Middlel-Middle Ages Ages site probable site funeraire isolated discovery environments. Nevertheless, because of the climate characteristics of the Sarliève environment, lower rates could be expected. This B % m suggests that SY rates calculated from Sarliève sediment stores, 100 450 although minimum, are quite close to gross erosion rates. Sediment yield is mainly related to soil erosion. Assuming that 80 3 400 the mean volumetric mass of eroded soils is 2 g/cm , it is possible to 60 converted the sediment yield (in t/km2 per yr) in catchment-soil ero- 40 sion rates (in mm/yr). During SYp1 to 5 as there is no colluvium of 350 this time on slopes and SY calculated values are close to true sedi- 20 ment flux values, we assumed that SY rates correspond to erosion 0 300 rates. During SYp6 and 8, because of colluvium stores on slopes and matter delivery downstream the lake, erosion rates deduced from e-Iron l-Empire e-Bronze l-Bronze e-Empire l-Neolithic m-Bronze e-La Tène l-La Tène SSY rates are evidently underestimated, especially DSY. m-Neolithic m-La Tène e-Middle Agesl-Middle Ages During the Lateglacial and Holocene, TSY, SSY and DSY
elevation (m) >500 400-500 350-400 <350 varied greatly (Figure 6). Examination of archaeological and paly- nological data provides clear indication of two periods in the mean elevation (m) evolution of the palaeoenvironment: the first before 7500 yr cal. BP during which SY was due only to natural bioclimatic condi- C ha tions (SYp1 to 4), and the second, after this date, during which SY 50 was also due to human impact (SYp5 to 9). 40
30 Sediment yield evolution before evidence of human impact 20 (Lateglacial and early Holocene) 10 While there is evidence of human settlement in the Sarliève catch- 0 ment prior to the Neolithic, it is sparse (Pomel, 1853; Pommerol, 1877; Trément et al., 2007b). Palynological data show no evidence e-Iron e-Bronze l-Bronze e-Empirel-Empire of farming before the Atlantic (Gachon, 1963): cereals only appear l-Neolithic m-Bronze e-La Tène l-La Tène e-Neolithicm-Neolithic m-La Tène e-Middlel-Middle Ages Ages in pollinic zone b (Figure 7A). Although ruderals were present scattered sites total sites from the beginning of the Atlantic, it is likely that variations in vegetation and hydrological processes influencing erosion prior to Figure 9. Stastistics of archaeological sites by period. e, early; about 7500 yr cal. BP were only due to natural processes and cli- m, middle; l, late. (A) Site numbers; (B) abundance of sites as mate change. During the Lateglacial and early Holocene, TSY was per cent of elevation class, and mean elevation of sites in metres; 2 (C) cumulated surface area of sites in ha lower (mean 55 t/km per yr) than subsequently, but variable (Figures 6A and 10): from 6 to 65 t/km2 per yr for SSY and 4 to 31 t/km2 per yr for DSY, with a DSY/SSY ratio of about 0.5. rich in organic matter containing archaeological artefacts (Vernet et al., 2005) and clearly the result of Holocene soil erosion, is Sediment yield phase 1 (SYp1) mainly located on the edge of the palaeolake. As it is not wide- During SYp1 (Bölling and beginning of the Alleröd), very low spread in the catchment, it was not quantified. SSY and DSY rates could be due to the first postglacial develop- Evaluating real DSY rates is difficult because of high discharge ment of arboreal cover with Juniperus and Betula (de Beaulieu of dissolved matter downstream of the lake during high-water et al., 1988) which limited mechanical and chemical erosion periods (Fourmont et al., 2009). DSY rates during the Lateglacial during a cool period (Johnsen et al., 1992 ). and early Holocene (SYp1 to SYp4) are probably close to gross rates resulting from the endorheic conditions of the palaeolake Sediment yield phase 2 (SYp2) (Einsele and Hinderer, 1998). During the middle and particularly Increased SSY (65 t/km2 per yr) during the Alleröd at 13 700 yr cal. the late Holocene, because the basin was open more often, first for BP, despite the continuation of previous bioclimatic conditions, Macaire et al. 509
Sediment yield rates Chronology Catchment surface Climate characteristics A B CD EF GH IJ KL M
SY Dissolved phases Solid Ages (yrs cal BP)ChronozonesArchaeological Lithology Tertiary Site surface SYp9 0 phases and non-arboreal vegetation SYp8 b 1000 Middle- Ages Emp SYp7 a 2000
sub-Atlantic Iron 3000 SYp6
4000 Bronze
5000 sub-Boreal
6000
SYp5 Neolithic 0 +2°C wet dry 7000
8000 Atlantic 0 50 100 0 20 40 mainly carbonated rocks SYp4 0 20 40 9000
Ruderals 10000 Boreal Tertiary b taxa Poaceae and cerals SYp3 a 11000 Pre boreal 12000 –4 –2 0°C tephr a SYp2 Yr . Dryas 13000
SYp1 14000 Bölling- Allerö d 15000 0 50 100 150 200 0 50 0.1 1 10 100 −8 −6 −4 –2 t.km−2.yr−1 t.km−2.yr−1 ha.10 yrs δ18O (‰)
Figure 10. Comparison of sediment yield rates, chronology, catchment area characteristics and climate. (D) Ages in yr cal. BP; (H) tertiary taxa as per cent of total pollen grains; Poaceae, and ruderals and cereals as per cent of catchment-taxa pollen-grains; (J) temperature from Greenland ice-core (Johnsen et al., 2001); (K) temperature from pollen data (Davis et al., 2003); (L) humidity from bog (Barber et al., 1994); (M) temperature from lake record (von Grafenstein et al., 1999)
seems to be due to CF1 tephra fallout, with a mean depth of 15 cm (Johnsen et al., 2001; Davis et al., 2003) (Figure 10K,M). Gachon (Vernet and Raynal, 2000). Tephric cover could have favoured (1963) suggested that the abundance of Chenopodiaceae pollen mechanical erosion over several decades or centuries (Figure 10G) grains in Sarliève sediments during the Lateglacial and early because of the temporary destruction of vegetation dominated by Holocene could have been due to slope-mass movement by soli- Betula and Pinus on slopes (de Beaulieu et al., 1988), but mainly fluction, which could have contributed to the development of these because of the high susceptibility of tephra to mechanical erosion pioneering taxa. The abundance of Chenopodiaceae at the begin- (Chorley et al., 1984; Collins and Dunne, 1986). Likewise, ning of the Atlantic (zone a, Figure 7C), when SSY was very low increased DSY (31 t/km2 per yr) can be explained by high suscep- (6 t/km2 per yr), indicates that there is no relationship between the tibility to dissolution of tephra silicates, mainly glass (Fourmont abundance of this taxa and the intensity of mechanical erosion. On et al., 2006). These higher SSY and DSY rates continued during the the other hand, Chenopodiaceae could have colonized the bottom Younger Dryas up to about 11 500 yr cal. BP, although tephric of the lacustrine depression during drought periods, providing cover had probably been completely eroded by that time; Tertiary evidence of the presence of salt water as shown elsewhere by pollen-grain abundance in lacustrine sediments (Gachon, 1963) Valero-Garcés et al. (2000). Salt, also testified by precipitated shows that mechanical erosion mainly cut into the Oligocene bed- minerals in B1 to B3 lacustrine units (Fourmont et al., 2006; rock. During the cold and dry Younger Dryas (Peteet, 1995; Bréhéret et al., 2008), was produced by dissolution of gypsum Anderson, 1997) (Figure 10M) mechanical erosion was favoured forming a small part of Oligocene limestones and marls (Jeambrun by tundra (de Beaulieu et al., 1988), which does not provide effi- et al., 1973). Dissolution of catchment bedrock gypsum and calcite cient protection of the earth surface against erosion, as observed in during SYp3 could indicate the beginning of pedogenesis during the Lac Chambon catchment located in a granitic and volcanic area the early Holocene and is in accordance with a relatively high close to the Sarliève catchment (Macaire et al., 1995, 1997). value of DSY (31 t/km2 per yr). However, it was not possible to Nevertheless, high DSY indicates that the climate was fairly evaluate precisely the evolution of DSY rates from SYp2 to SYp3. humid, and the dissolution of carbonated rocks could have been favoured by the cold climate. Sediment yield phase 4 (SYp4) During the Boreal and the Atlantic before 7500 yr cal. BP, erosion Sediment yield phase 3 (SYp3) rates, much lower than before and similar to SYp1, did not change At the beginning of the Holocene, between 11 500 yr cal. BP and significantly (SSY = 6 to 11 t/km2 per yr and DSY = 4 to 9 t/km2 9800 yr cal. BP, SSY and DSY remained high, although the former per yr), although the DSY/SSY ratio increased (0.36 to 1.50; fell to 46 t/km2 per yr (Ep3b). At this time, forest developed, Figure 6D). This trend could be explained by extensive develop- dominated first by Corylus and then Quercus and Ulmus (Gachon, ment of forest cover, dominated by Corylus and Quercus (de 1963; de Beaulieu et al., 1988), and the climate became warmer Beaulieu et al., 1988; Gachon, 1963), related to climate warming 510 The Holocene 20(4)
(Figure 10K,M), which provided efficient protection of the earth simultaneity between periods of deforestation and periods of high surface against mechanical erosion. Higher temperatures and less Tertiary pollen grain content (Figure 7B) and detrital layer deposi- precipitation during the Boreal, testified in western Europe tion (HCS facies in B3), indicating increased mechanical erosion. (Andrews, 2006) with minimum temperatures and dryness at In the lacustrine depression during SYp5, a marked variation of about 8200 yr cal. BP in the Northern Hemisphere (Carrion, 2002; arboreal pollen contents (mainly Alnus) compared to Cyperaceae Alley and Agustsdottir, 2005; Macklin et al., 2006), may have and paludals (Figure 7D), appears to be related to an alternation of been more marked in the drier microclimate of Sarliève and could biochemical dominated BL facies and low water level in the pal- explain the low absolute value of DSY. Nevertheless, the sharp aeolake on the one hand, and detrital dominated HCS facies and rise of the DSY/SSY ratio at this period, marked by catchment- high water level observed in lacustrine sediments on the other slope stability, indicates a matter budget promoting an increase in (Fourmont et al., 2009). Non-pollinic microfossils indicate the soil depth. occurrence of livestock farming as early as pollinic zone c, which could explain the mesotrophic to eutrophic character of the pal- aeolake waters (Argant and Lopez-Saez, 2004). These marked Sediment yield evolution during periods of variations in vegetation, hydrology and sediment supply in the human impact (middle and late Holocene) lacustrine depression are similar to variations in the catchment TSY increased dramatically after 7500 yr cal. BP (average for vegetation. Low water level in the lake could be the result of SYp5, SYp6 and SYp8: 165 t/km2 per yr): from 24 to 232 t/km2 per increased evapotranspiration during afforestation in the catch- yr for SSY and from 6 to 44 t/km2 per yr for DSY. The DSY/SSY ment, and high water level due to increased runoff during defores- ratio decreased to close to 0.25, with considerable fluctuations tation. However, the relative impacts of climate versus human during the Atlantic (SYp5) (Figure 6). This rise in sediment yield activity on hydrology and sediment yield are not yet clear because rate appeared concomitantly with the growth of human settlement of the marked variability of sedimentary facies in unit B3. High in the Sarliève catchment (Figure 10I). Pottery dated from the end resolution multiproxy analysis (at a centimetric scale) of this unit of the Ancient Neolithic, although scarce (two probable sites), is in progress. provides evidence of human activity in the catchment; similar activity has also been observed in many places around Clermont- Sediment yield phase 6 (SYp6) Ferrand from 7500 yr cal. BP (Georjon et al., 2004; Pouenat, From the sub-Boreal after 5300 yr cal. BP up to 2300 yr cal. BP in 2007). Human activity increased considerably after that period, the sub-Atlantic, increased TSY (201 t/km2 per yr) and rising with some fluctuations (Figures 8 and 9). Evidence of ruderals water level (Fourmont et al., 2009) are in keeping with changes in (zone a, Figure7A) and cereals (zone b, Figure 7A) from the the climate which became more humid between 5200 yr cal. BP beginning of the Atlantic show that vegetation and earth surface and 4500 yr cal. BP (Barber et al., 1994; Magny et al., 2006) and conditions were modified by human activities from about 7500 yr cooler (Johnsen et al., 2001), with greater cooling at about 2800 BP. Human activity and climate together influenced erosion yr cal. BP (van Geel et al., 1998) (Figure 10J,L). Nevertheless, a processes and intensity (Figure 10H–L). greater rise in SSY and DSY during SYp6 than SYp5 (by factors of 2.9 and 6.6, respectively, Figures 6B and C) cannot be Sediment yield phase 5 (SYp5) explained solely by climate change which was not marked. The From the Atlantic after 7500 yr cal. BP to the beginning of the rise in SSY is synchronous with increased human settlement sub-Boreal at about 5300 yr cal. BP, sharp TSY increase (61 t/km2 between the end of the Neolithic and the ancient La Tène as shown per yr) is characterized by strong variations in DSY/SSY ratio. by the abundance of sites, particularly during the ancient Bronze Marked SSY increase (by a factor of nine maximum relative to Age, final Bronze Age and first Iron Age (Figures 8B and 9A): SYp4) together with lower variation in DSY (increase by a factor total surface area of sites varies between archaeological periods of four maximum) is an indication of human impact (Dearing and (Figure 9C), but the mean time-weighted site area increased more Jones, 2003), although climate became a little more humid than during SYp6 (0.92 ha/century) than SYp5 (0.65) (Figure 10I). before. SYp5 corresponds to the development of the Neolithic These sites, located at elevations between 350 and 500 m from the society and arable farming in the Massif Central (Georjon et al., final Neolithic until the middle La Tène (Figure 9C), were scat- 2004), particularly widespread in the Sarliève catchment during tered between 700 m and the edge of the palaeolake (< 350 m) the Chassean Middle Neolithic (Figures 8A, 9A). Evidence of from the final Bronze until the middle La Tène, promoting transit farming is provided by cereal pollen grains (Figure 7A). Farming of solid flux from top to bottom of slopes. During SYp6, forest seems to have occurred mainly at elevations of 350 to 500 m cover on slopes, marked by the replacement of Corylus by Fagus (average 400–450 m, Figure 9B) where ancient and middle (zones e to h, Figure 7A) was more continuous than during SYp5. Neolithic sites cover a total area of about 13 ha (Figure 9C), cor- Nevertheless, several maximum cereal-pollen contents, often syn- responding to a mean time-weighted settlement area of 0.65 ha per chronous with increased Tertiary taxa and slight forest regression century (Figure 10I). It is difficult to explain high variations of (Figure 7A, B), show that human impact and mechanical erosion arboreal pollen grain contents in slope vegetation during the of bedrock could have been intense during certain periods (pol- Atlantic (zones a, b and c Figure 7A) by climate change, which, linic zone e in particular). These periods could correspond to rapid although not marked, did occur (Aaby, 1976; Bond et al., 2001; human population growth, as during the ancient Bronze Age (pol- Douglas et al., 2007). On the other hand, a sharp increase in ruder- linic zone e) and ancient Iron Age (zone g) and explain the high als (Figure 7A) and preferential decrease in Corylus (Figure 11B), overall SSY value during SYp6. The palaeolake water level, which which is more easily cut than Quercus (Figure 11A), during peri- was probably higher because of the more humid climate, varied as ods of slope arboreal pollen regression, could indicate human shown by non-pollinic microfossils (Argant and Lopez-Saez, impact on deforestation. This seems to be confirmed by frequent 2004). At the edge of the palaeolake, where vegetation varied Macaire et al. 511
A sediment stores dating from these periods. Archaeological evidence 30 such as drainage ditches, buildings, funerary sites, and off-site pottery for soil enrichment observed at the bottom of the lacustrine 2 R = 0.09 depression show that it was dry and mainly settled by humans at 20 the end of the La Tène and during the early Roman Empire (Figure 8C). These dry conditions could have been partly due to a less humid climate at this time (Barber et al., 1994; Magny, 2004) 10 (Figure 10L), but were mainly the result of human activity as Pollens of Quercus shown by drainage ditches and parcels with boundary markings in as % of total pollen grains
0 the depression (Trément et al., 2007b). A marked rise in settle- 0 20 40 60 80 100 ments on catchment slopes is shown by very abundant sites of the Pollens of catchment trees as % of total pollen grains La Tène and early Roman Empire (Figure 9A), especially at low elevations (Figure 9B), and the high value of time-weighted sur- B 60 face area of sites (21 ha/century) (Figures 9C and 10I). Archaeological data (Trément et al., 2007b) and the increasing 50 extent of enriched soils (Figure 8C) indicate intense arable farm- 2 R = 0.60 ing, suggesting a high erosion rate. The lack of deposits is due to 40 the discharge of most solid and dissolved matter downstream of 30 the catchment through efficient drainage ditches.
20 Sediment yield phase 8 (SYp8)
Pollens of Corylus 10 During the sub-Atlantic, from 1800 yr cal. BP until 300 yr cal. BP, 2 as % of total pollen grain s TSY (222 to 247 t/km per yr) increased slightly relative to SYp6 0 (by a factor of 1.1 to 1.2) (Figure 6A). After the great development 0 20 40 60 80 100 Pollens of catchment trees as % of total pollens of the final La Tène and early Roman Empire societies (SYp7), similar SY rates could be explained by a sharp reduction in the number and total surface area of sites (time-weighted surface area Figure 11. Correlations between catchment arboreal taxa from total = 3.5 ha/century) during the late Roman Empire and early Middle pollen grain percentages. (A) Quercus versus catchment trees; Ages (Figures 8D and 9A,C), down to rates close to those for (B) Corylus versus catchment trees SYp6. Nevertheless, during SYp8, there was a marked decrease in Pinus-dominated forest on slopes (less than 50% of arboreal pol- more than on catchment slopes (Figure 7D), periods of Cyperaceae len grains, Figure 7A). Moreover, the increase in Chenopodiaceae and paludal increase and arboreal taxa decrease can be explained initiated during the La Tène (Figure 7C) cannot be attributed to by frequent variations in water level, recently emerged areas being water salinity, as during the Lateglacial and early Holocene, colonized by hygrophilous herbs. Nevertheless, Cyperaceae prob- within the context of high water levels and greater humidity, par- ably indicate above all the occasional but increasing deforestation ticularly during the last millenium (Barber et al., 1994; of the edge of the palaeolake by humans during SYp6, especially Magny, 2004; Magny et al., 2008) (Figure 10L); this taxa, with the during the ancient La Tène (Figure 9B,C, and zone h Figure 7D), increase in other ruderals, more likely indicates a strong anthropo- a period of extensive human settlement confirmed by increasing genic impact on vegetation. The presence of pyrofusinite-type eutrophication of palaeolake water and evidence of livestock charcoal fragments (Fourmont, 2005) and fungi growing on farming (Argant and Lopez-Saez, 2004). Periods of high defores- charred tissue (Argant and Lopez-Saez, 2004) in black clayey- tation round the palaeolake do not seem to be synchronous with silty B5 unit, are also evidence of frequent fires during the late that on catchment slopes (Figure 7A, D), indicating possible dif- Roman Empire, probably for soil enrichment rather than due to ferences in land use from one area to another; as a result, matter forest clearing for farming which was already extensive. The produced by mechanical erosion could have moved spasmodically deforested area was first colonized by grassland (zone 1, Figure toward the lake. 7A) for livestock as shown by coprophilous fungi and evidence of Although calcite dissolution is favoured by a humid climate, palaeolake eutrophication. Despite this high anthropogenic impact the very high increase in DSY during SYp6 cannot be explained on vegetation, low SSY increase during the late Roman Empire only by climate change. Dissolved matter is mainly composed of (SYp8a) could be explained by the development of grassland at carbonate ions because of weathering of marly-calcareous bed- different slope elevations (Figure 9B) down to the edge of the rock. It is probable that strong soil rejuvenation by mechanical palaeolake where trees almost disappeared (Figure 7D); pasture- erosion reactivated the weathering of parent-rocks, and thus DSY, land provides efficient protection of the earth surface against as suggested by Carson and Kirkby (1972) and observed in the mechanical erosion, except when overgrazed (Ursic and Dendy, nearby Chambon catchment by Gay and Macaire (1999). Low 1965; Pimentel et al., 1995; Hooke, 2000). DSY/SSY ratio (# 0.25) expresses this imbalance in soil evolution. From the beginning of the Middle Ages, croplands developed on catchment slopes, as shown by increased cereal and fallow-land Sediment yield phase 7 (SYp7) taxa (zone j, Figure 7A), while palaeolake eutrophication was at its It is not possible to evaluate sediment yield rates during SYp7, highest (Argant and Lopez-Saez, 2004). Settlement sites, less corresponding to the middle and final La Tène and the early numerous and extended than before, were concentrated around the Roman Empire (2300–1800 yr cal. BP), because there are no palaeolake (Figures 8D and 9B). During SYp8b (the Middle Ages 512 The Holocene 20(4)
and Modern Times), high SSY (203 t/km2 per yr) mainly due to (by factors of > 38 and > 11, respectively), while they varied less mechanical erosion of sandy-gravelly sediments from the alluvial (by factors of 11 and 8, respectively) during the Lateglacial and terrace east of the palaeolake (Figure 1), which produced D8 del- early Holocene (SYp2–SYp4) which was marked by considerable taic unit, could be the result of agricultural concentration in this climate change. Thus, increased SYrates during SYp6 and SYp8, area. Moreover, colluvium overlying archaeological sites of the late particularly SSY, seem to be highly related to the growth of Roman Empire and early Middle Ages at the slope bottom indicates human activity. increased mechanical erosion after these periods. This colluvium SSY (174 t/km2 per yr on average for SYp6 and SYp8) rose by storage could have been promoted by temporary afforestation a factor of 6.4 (i.e. 540%) during the late Holocene compared with (Alnus) round the palaeolake (first part of zone j, Figure 7D). These earlier Holocene periods (27 t/km2 per yr on average for SYp3 to colluvium stores were not taken into account when quantifying SYp5), while precipitation is assumed to have risen by 50 mm to SYp8b SSY which is thus underestimated. Rates of DSY (mini- 100 mm (Guiot et al., 1989), i.e. about 10%, during the sub-Boreal mum of 44 t/km2 per yr) and the DSY/SSY ratio (# 0.25) do not and sub-Atlantic. Any increase in rainfall may directly exacerbate differ much from SYp6 rates, showing similar pedogenetic condi- erosion (Boardman, 1990; Bullock, 1991), but the relationship tions with high soil rejuvenation by mechanical erosion. between these two factors is not linear: a 10% increase in winter rainfall can induce an increase in mechanical erosion of up to 150% Sediment yield phase 9 (SYp9) (Favis-Mortlock and Boardman, 1995). Nevertheless, this assess- SY cannot be evaluated for the last three centuries as sediment ment concerns currently tilled soils which are particularly suscep- yield has been discharged downstream of the Sarliève catchment tible to mechanical erosion, while precipitation at the beginning of through ditches, called ‘rases’, since the palaeolake was drained the Holocene rose in extensively forested environments. SSY does during the seventeenth century (Fournier, 1996). not seem to have increased by more than 150% as a result of rising precipitation and erosion, taking all other environmental parame- ters as constant. SSY could have increased by at least 390%, i.e. Balance of climate change versus human 75% of total increase, as a result of human activity during the final impact on sediment yield, earth surface Neolithic, probably exacerbated by climate change. SSY increased lowering and soil evolution more in SYp5 than SYp6, and less in SYp6 than SYp8, with a break During the Lateglacial and early Holocene (SYp1 to SYp4), SSY in the intensity of erosion during the final Neolithic, which is a and DSY rates, lower on average than later, depended mainly on little earlier than the break observed by Miras et al. (2004) in the strong climate change, from cold to warm, with a general trend higher Massif Central. As there are few final Neolithic sites at towards dryness (Figure 10). Highest SSY rates (SYp2 and SYp3) Sarliève, this break could be a delayed effect of human settlement correspond to the coldest periods (essentially the Younger Dryas during the middle Neolithic, as observed elsewhere (Edwards and and Preboreal), while lowest rates (SYp1 and SYp4) correspond to Whittington, 2001); forest cover was still abundant at this time and vegetation development during warmer periods (Bölling-Alleröd could have slow down the solid flux delivery to the palaeolake. Soil and Boreal). These erosion trends are similar to those observed in use changes during the Protohistorical and Historical periods do not the Lac Chambon catchment (Macaire et al., 1997; Gay and seem to have significantly affected mechanical erosion intensity Macaire, 1999), located close to the Sarliève catchment but at a before the seventeenth century. higher elevation and in a crystalline bedrock area, confirming DSY increase during the late Holocene is more difficult to regional climate impact on sediment yield. The trend in SSY val- explain, as their rates are certainly strongly underestimated for the ues at Sarliève differs from that observed in mountainous areas sub-Boreal and sub-Atlantic. Taking all other environmental param- where abundant inherited glacial deposits induced high SSY after eters as constant, a 10% increase in water infiltration into the soil glacier retreat (Church and Ryder, 1972; Bichet et al., 1999; and a 1.5°C drop in the mean temperature at about 4500 yr cal. BP Hinderer, 2001). However, in the Sarliève catchment, climatic could have led to an increase in chemical erosion (calcite dissolu- impact on SYwas disturbed by CF1 tephra fallout at 13 700 yr cal. tion) of less than 15%. Thus, at least 90% of the DSY increase BP, leading to increased SSY from the Alleröd, before the climate (overall 180%) between SYp5 and SYp6 could also be anthropo- change of the Younger Dryas, confirming strong and short-term genic, increased SSY reactivating DSY. The impact of climatic impact of surface formations on mechanical erosion. The trend in versus anthropogenic factors on SY during SYp5, a period character- DSY values is fairly similar to SSY (Figure 10B and C), with an ized by very variable SSY and DSY rates and the development of increase during cold periods. However, temperature, relative to ancient and middle Neolithic society, is currently being investigated. precipitation, is not the main factor of carbonate dissolution From estimation of mean eroded soil thickness from SSY and (Gombert, 1997), and the similar DSY and SSY trends suggest DSY rates and duration of each SY phase, three distinct periods that mechanical erosion favoured chemical erosion through soil can be identified (Figure 12): (1) from the Bölling to the beginning rejuvenation, so that parent-rocks, which are more susceptible to of the Boreal (SYp1 to SYp3), total earth surface lowering in the dissolution, were found close to the earth surface (Carson and catchment was about 20 cm thick; soil thickness was low as shown Kirkby, 1972; Berner and Berner, 1987; Gabet and Mudd, 2009); by relatively constant and low DSY/SSY ratio (# 0.5) in environ- carbonated rocks, and temporarily tephra, were particularly reac- ments marked by cold to cool and dry climate and sparse vegeta- tive to this process. tion; (2) during the Boreal and Atlantic, during which erosion rates During the late Holocene (SYp6 and SYp8), increasing SSY were probably still close to SY rates, total erosion was low (8 cm and DSY are in keeping with lower temperatures and rising pre- thick) with rise in soil thickness shown by high DSY/SSY ratio cipitation after 5200 yr cal. BP (Johnsen et al., 2001; Barber et al., (up to 1.5) as a result of vegetation cover growth and warmer cli- 1994). However, SSY and DSY during the Holocene (SYp4–SYp8), mate, although the lower ratios during SYp5 could indicate the when there was no marked climate change, varied extensively beginning of anthropogenic impact; (3) during the sub-Boreal and Macaire et al. 513
Earth surface lowering Possible Without human impact human impact With human impact (in mm) SYp1 SYp2 SYp3 SYp4 SYp5 SYp6 SYp7 SYp9 SYp8 0
100
200
300
due to human impact 400
500
600
700
Beginning of soil formation Soil thickening Soil rejuvenation 800
Bölling Alleröd Younger pre- sub-Boreal sub-Atlantic Dryas Boreal Atlantic Boreal 900 14000 12000 10000 8000 6000 4000 2000 0 ages in yrs cal BP calculated mechanical erosion assumed maximal mechanical erosion due to climate change calculated minimal total erosion (mechanical + chemical) assumed minimal total erosion due to climate change
Figure 12. Evaluation of earth surface lowering and soil thickness from sediment yield rates since the Lateglacial sub-Atlantic, soil erosion rates were much higher than SY rates: did not significantly alter SYrates. Extensive vegetation clearing total erosion greatly increased (55 cm thick at least) with extensive due to climate change during the Younger Dryas and to human soil rejuvenation and thinning characterized by a pronounced low- activity during the Historic Times can explain periods of severe ering of the DSY/SSY ratio (# 0.25). This trend appears to be a mechanical erosion at some places (delivery of coarse-grained del- consequence of human activity which is assumed to have taic sediments) during SYp2 and SYp8. Catchment-soil evolution accounted for more than 75% of SSY and 90% of DSY, compared deduced from DSY and SSY rates and DSY/SSY ratio indicate that with SY rates resulting only from climate during the late Holocene. soils began to be formed during the Lateglacial and Preboreal, then thickened considerably during the Boreal and Atlantic, and finally became thinner (rejuvenation) mainly as the result of human- Conclusion induced erosion during the sub-Boreal and sub-Atlantic. Severe By quantifying stored sediment mass in the Sarliève palaeolake it mechanical erosion during the late Holocene induced a sharp rise in was possible to calculate minimum solid (SSY) and dissolved chemical erosion, which was favoured by the presence of mainly (DSY) sediment yield rates for seven phases (SYp1 to SYp6 and carbonated rocks in the catchment. SYp8) between about 15 000 yr cal. BP and 300 yr cal. BP; phases SYp7 and SYp9 were marked by anthropogenic drying of the lake and lack of sediment stores. During the Lateglacial and early Acknowledgments Holocene up to 7500 yr cal. BP, SSY and DSY rates, lower on aver- This work was supported by the French CNRS-INSU ECLIPSE age than later, varied mainly in relation to climate change: higher programs ‘Interactions activités humaines – production et stock- rates (SYp2 and SYp3) correspond to colder periods (Younger age de sédiments à l’Holocène en plaine et en moyenne mon- Dryas and Preboreal), while lower rates (SYp1 and SYp4) corre- tagne’ and Zone Atelier Loire. We thank J.P. Bakyono, I. Pène and spond to periods of warmer climate (Bölling-Alleröd and Boreal). many other collaborators for their help in acquiring field data and However, CF1 tephra fallout at 13 700 yr cal. BP induced greater sediment analysis, and the Institut National de Recherche SY during the Alleröd. After 7500 yr cal. BP, the middle and late d’Archéologie Préventive (INRAP) and the Society Fondasol Holocene are marked by a sharp increase in SSY and DSY, reaching Auvergne for their collaboration. We thank Miss Yates for their a peak during the final Neolithic after 5300 yr cal. BP. Comparison assistance in translation of this text and two anonymous reviewers of SY rates with quantified data of human settlement, and palaeo- for their constructive comments. climatic and palynological data, indicates that after 5300 yr cal. BP, when the climate became cooler and more humid, at least 75% of SSY increase and more than 90% of DSY increase were due to References human activities. Changes in soil use during the Protohistoric Times Aaby, B. 1976: Cyclic climatic variations in climate over the past (SYp6) and the Historic Times up to the seventeenth century (SYp8) 5,500 yr reflected in raised bogs. Nature 263, 281–84. 514 The Holocene 20(4)
Alley, R.B. and Agustsdottir, A.M. 2005: The 8ka event: cause and Carrion, J.S. 2002: Patterns and processes of Late Quaternary consequences of a major Holocene abrupt climate change. environmental change in a montane region of southwestern Quaternary Science Reviews 24, 1123–49. Europe. Quaternary Science Reviews 2, 2047–66. Anderson, D.E. 1997: Younger Dryas research and its implications Carson, M.A. and Kirkby, M.J. 1972: Hillslope form and process. for understanding abrupt climatic change. Progress in Physical Cambridge University Press, 475 pp. Geography 21, 230–49. Cerdan, O., Poesen, J., Govers, G., Saby, N., Le Bissonnais, Y., Andrews, J.E. 2006: Palaeoclimatic records from stable isotopes in river- Gobin, A., Vacca, A., Quinton, J., Auerswald, K., Klik, A., ine tufas: Synthesis and review. Earth-Science Reviews 75, 85–104. Kwaad, F.J.P.M. and Roxo, M.J. 2006: Sheet and rill erosion. Antoine, P. and Fabre, D. 1980: Géologie appliquée au génie civil. Chapter 2.4. In Boardman, J.and Poesen, J., editors, Soil erosion Masson éd., Paris, 291 pp. in Europe. Wiley, 501–13. Argant, J. and Lopez-Saez, J.-A. 2004: L’occupation humaine du Chorley, R., Schumm, S.A. and Sugden, D. 1984: Geomorphology. bassin de Sarliève depuis le Néolithique d’après la palynolo- Methuen, 689 pp. gie et l’étude des microfossiles non polliniques. 5e Rencontres Church, M. and Ryder, J.M. 1972: Paraglacial sedimentation: a consid- méridionales de Préhistoire récente Auvergne et Midi, Actes de eration of fluvial processes conditioned by glaciation. Geological la cinquième session, Clermont-Ferrand, et 9 novembre 2002, Society of America Bulletin 83, 3059–72. Préhistoire du Sud-ouest. Suppl. 9, 25–34. Collins, B.D. and Dunne, T. 1986: Erosion of tephra from the 1980 Ballut, C. 2000: Evolution environnementale de la Limagne de eruption of Mount St. Helens. Geological Society of America Clermont-Ferrand au cours de la seconde moitié de l’Holocène Bulletin 97, 896–905. (Massif central français). PhD thesis, University Blaise Pascal of Davis, B.A.S., Brewer, S., Stevenson, A.C., Guiot, J. and Data Clermont-Ferrand, 315 pp. Contributors 2003: The temperature of Europe during the Barber, K.E., Chambers, F.M., Maddy, D., Stoneman, R. and Brew, Holocene reconstructed from pollen data. Quaternary Science J.S. 1994: A sensitive high-resolution record of late Holocene Reviews 22, 1701–16. climatic change from a raised bog in northern England. The Dearing, J.A. and Jones, R.T. 2003: Coupling temporal and spatial Holocene 4, 198–205. dimensions of global sediment flux through lake and marine Berner, E.K. and Berner, R.A. 1987: The global water cycle. Prentice- sediment records. Global and Planetary Change 39, 147–68. Hall, Inc., 397 pp. de Beaulieu, J.L., Pons, A. and Reille, M. 1988: Histoire de la flore Bichet, V., Campy, M., Buoncristiani, J.F., Di-Giovanni, C., Meybeck, et de la végétation du Massif Central (France) depuis la fin de la M. and Richard, H. 1999: Variations in sediment yield from the dernière glaciation. Cahier de Micropaléontologie 3, 5–36. Upper Doubs River carbonate watershed (Jura, France) since the Douglas, I. 1967: Man, vegetation and the sediment yields of rivers. Late-Glacial Period. Quaternary Research 51, 267–79. Nature 215, 925–28. Boardman, J. 1990: Soil erosion on the South Downs: a review. In Douglas, R., Gonzalez-Yajimovich, O., Ledesma-Vazquez, J. and Boardman, J., Foster, I.D.L. and Dearing, J.A., editors, Soil Staines-Urias, F. 2007: Climate forcing, primary production and erosion on agricultural land. Wiley, 87–105. the distribution of Holocene biogenic sediments in the Gulf of Boivin, P., Besson, J.C., Briot, D., Camus, G., de Goër de Herve, California. Quaternary Science Reviews 26, 115–29. A., Gourgaud, A., Labazuy, P., de Larouzière, F.D., Livet, M., Edwards, K.J. and Whittington, G. 2001: Lake sediments, erosion and Mergoil, J., Miallier, D., Morel, J.M., Vernet, G. and Vincent, landscape change during the Holocene in Britain and Ireland. P.M. 2004: Volcanologie de la Chaîne de Puys. Ed. Parc Régional Catena 42, 143–73. des Volcans d’Auvergne, 179 pp. Einsele, G. and Hinderer, M. 1998: Quantifying denudation and Bond, G., Kromer, B., Muscheler, R., Evans, M.N., Showers, W., sediment-accumulation systems (open and closed lakes): basic Hoffmann, S., Lotti-Bond, R., Hajdas, I. and Bonani, G. 2001: concepts and first results. Palaeogeography, Palaeoclimatology, Persistent solar influence on North Atlantic climate during the Palaeoecology 140, 7–21. Holocene. Science 294, 2130–36. Favis-Mortlock, D. and Boardman, J. 1995: Nonlinear responses of Bréhéret, J.G., Macaire, J.J., Fleury, A., Fourmont, A. and Soulié- soil erosion to climate change: a modelling study on the UK Marsche, I. 2003: Indices de confinement dans les dépôts lacus- South Downs. Catena 25, 365–87. tres holocènes de Sarliève (Limagne, France). Comptes Rendus Fourmont, A. 2005: Quantification de l’érosion et de la sédimenta- Géosciences 335, 479–85. tion dans le bassin de Sarliève (Massif central, France) au Bréhéret, J.G., Fourmont, A., Macaire, J.J. and Négrel, Ph. 2008: Tardiglaciaire et à l’Holocène. Impact des facteurs naturels et Microbially mediated carbonates in the Holocene lacustrine anthropiques. PhD thesis, University of Tours, 419 pp. deposits from Sarliève, a small ancient lake of the French Massif Fourmont, A., Macaire, J.-J., Bréhéret, J.-G., Argant, J., Prat, B. and Central, testify to the evolution of a restricted environment. Vernet, G. 2006: Tephras in lacustrine sediments of the Sarliève Sedimentology 55, 557–78. marsh (French Massif Central): age and preservation. Comptes Brune, G.M. 1953: Trap efficiency of reservoirs. American Rendus Géosciences 338, 1141–49. Geophysical Union Transactions 34, 407–18. Fourmont, A., Macaire, J.-J. and Bréhéret J.-G. 2009: Contrasted late Bullock, P. 1991: Soils. In United Kingdom Climate Change Impacts glacial and Holocene hydrology of Sarlieve paleolake (France) Review Group, editors, The potential effects of climate change in from sediment geometry and detrital vs. biochemical composition. the United Kingdom. HMSO, 15–25. Journal of Paleolimnology 41, 471–90. Campy, M., Bichet, V., Di-Giovanni, C., Richard, H., Richard, J. and Fournier, G. 1996: Sarliève: un lac au moyen âge. Association du Site Olive, P. 1994: Evolution des flux de matière depuis 12000 ans de Gergovie 11, 2–34. dans la haute vallée du Doubs (France). Bulletin de la Societe Gabet, E.J. and Mudd, S.M. 2009: A theoretical model coupling chem- Géologique de France 165, 381–400. ical weathering rates with denudation rates. Geology 37, 151–54. Macaire et al. 515
Gachon, L. 1963: Contribution à l’étude du quaternaire récent de la Un exemple, le bassin du Lac Chambon (Massif Central, France). Grande Limagne marno-calcaire: morphogenèse et pédogenèse. Comptes Rendus de l’Academie de Science Paris 320, 579–85. Annales agronomiques 14, 191 pp. Macaire, J.-J., Bossuet, G., Choquier, A., Cocirta, C., De Luca, P., Gay, I. and Macaire, J.-J. 1999: Estimation des taux d’érosion Dupis, A., Gay, I., Mathey, E. and Guenet, P. 1997: Sediment chimique tardiglaciaires et holocènes par la méthode des bilans yield during Lateglacial and Holocene periods in the Lac d’altération. Application au bassin du lac Chambon (Massif Chambon watershed, Massif Central, France. Earth Surface Central, France). Comptes Rendus de l’Academie de Science Processes and Landforms 22, 473–89. Paris 328, 387–92. Macaire, J.-J., Bernard, J., Di-Giovanni, C., Hinschberger, F., Georjon, C., Jallet, F., Lagrue, A. and Loison, G. 2004: Le Néolithique Limondin-Lozouet, N. and Visset, L. 2006: Quantification ancien et moyen en Auvergne: bilan et perspectives à la lumière and regulation of organic and mineral sedimentation in a Late des données récentes. Actes des Ve Rencontres méridionales de Holocene floodplain as a result of climatic and human impacts Préhistoire récente, Clermont-Ferrand (8–9 novembre 2002). (the Taligny marsh, Parisian Basin, France). The Holocene 16, Préhistoire du Sud-Ouest. Suppl. 9, 107–32. 647–60. Goeury, C. and de Beaulieu, J.L. 1979: A propos de la concentration Macklin, M.G., Benito, G., Gregory, K.J., Johnstone, E., Lewin, J., du pollen à l’aide de la liqueur de Thoulet dans les sédiments Michczynska, D.J., Soja, R., Starkel, L. and Thorndycraft, V.R. minéraux. Pollens et spores 30, 73–89. 2006: Past hydrological events reflected in the Holocene fluvial Gombert, P. 1997: Quantification du rôle des paramètres climatiques record of Europe. Catena 66, 145–54. dans la dissolution karstique. Comptes Rendus de l’Academie de Magny, M. 2004: Holocene climate variability as reflected by Science Paris 324, 17–23. mid-European lake-level fluctuations and its probable impact Guiot, J., Pons, A., de Beaulieu, J.L. and Reille, M. 1989: A 140,000- on prehistoric human settlements. Quaternary International year continental climate reconstruction from two European pol- 113, 65–79. len records. Nature 338, 309–13. Magny, M., Leuzinger, U., Bortenschlager, S. and Haas, J.N. 2006: Heineman, H.G. 1984: Reservoir trap efficiency. In Hadley, R.F. and Tripartite climate reversal in central Europe 5600–5300 years Walling, D.E., editors, Erosion and sediment yield: some methods ago at Arbon-Bleiche, Switzerland. Quaternary Research of meseasurement and modelling. Geo Books, 201–18. 65, 3–19. Hinderer, M. 2001: Late Quaternary denudation of the Alps, valley Magny, M., Gauthier, E., Vannière, B. and Peyron, O. 2008: and lake fillings and modern river loads. Geodinamica Acta Palaeohydrological changes and human-impact history over 14, 231–63. the last millenium recorded at Lake Joux in the Jura Mountains, Hinschberger, F., Fourmont, A., Macaire, J.-J., Bréhéret, J.-G., Switzerland. The Holocene 18, 255–65. Guérin, R. and Bakyono, J.P. 2006: Contribution of geophysi- McNeil, J.D. 1980: Electromagnetic terrain conductivity mea- cal surveys to the study of fine grained lacustrine sediments. surements at low induction numbers. Technical note TN-6. Application to the Sarliève marsh (Massif Central, France). Geonics, 15 pp. Bulletin de la Societe Géologique de France 177, 311–22. Meybeck, M. 1979: Concentrations des eaux fluviales en éléments Hooke, R.L. 2000: On the history of humans as geomorphic agents. majeurs et apport en solution aux océans. Revue de Géologie Geology 28, 843–46. Dynamique et de Géographie Physique 21, 215–45. Jeambrun, M., Aubert, M., Bouiller, R., Camus, G., Cochet, A., —— 1987: Global chemical weathering of surficial rocks estimated from d’Arcy, D., Giot, D., Baudry, D., Roche, A. and Bonhommet, N. river dissolved loads. American Journal of Science 287, 401–28. 1973: Carte géologique à 1/50000 de Clermont-Ferrand XXV-31 Milliman, J.D. and Meade, R.H. 1983: World-wide delivery of river et notice. BRGM Ed. sediment to the oceans. Journal of Geology 91, 1–21. Johnsen, S.J., Clausen, H.B., Dansgaard, W., Fuhrer, K., Gundestrup, Milliman, J.D. and Syvitski, J.P.M. 1992: Geomorphic/tectonic N.S., Hammer, C.U., Iversen, P., Jouzel, J., Stauffer, B. and control of sediment discharge to the ocean: the importance of Steffensen, J.P. 1992: Irregular glacial interstadials recorded in a small mountainous rivers. Journal of Geology 100, 525–44. new Greenland ice core. Nature 359, 311–13. Miras, Y., Laggoun-Dégarge, F., Guenet, P. and Richard, H. 2004: Johnsen, S.J., Dahl-Jensen, D., Gundestrup, N., Steffensen, J.P., Multi-disciplinary approach to changes in agro-pastoral activi- Clausen, H.B., Miller, H., Masson-Delmotte, V., Sveinbjörndottir, ties since the Sub-Boreal in the surroundings of the ‘narse A.E. and White, J. 2001: Oxygene isotope and paleotemperature d’Espinasse’ (Puy de Dôme, French Massif Central). Vegetation records from six Greenland ice-core stations: Camp Century, History and Archeobotany 13, 91–103. Dye-3, GRIP, GISP2, Renland and NorthGRIP. Journal of Peteet, D. 1995: Global Younger Dryas? Quaternary International 28, Quaternary Science 16, 299–307. 93–104. Judson, S. and Ritter, D.F. 1964: Rates of regional denudation in Pimentel, D., Harvey, P., Resosudarmo, P., Sinclair, K., Kurz, D., United States. Journal of Geophysical Research 69, 3395–401. McNair, M., Crist, S., Shpritz, L., Fitton, L., Saffouri, R. and Kessler, J. and Chambraud, A. 1986: La météo de la France. Tous les Blair, R. 1995: Environmental and economic costs of soil erosion climats localité par localité. Lattès. and conservation benefits. Science 267, 1117–23. Ludwig, W. and Probst, J.L. 1998: River sediment discharge to the Pomel, M. 1853: Catalogue méthodique et descriptif des vertébrés fos- oceans: present-day controls and global budgets. American siles découverts dans le bassin hydrographique supérieur de la Journal of Science 298, 265–95. Loire et surtout dans la vallée de son affluent principal, l’Allier. Macaire, J.-J., Bossuet, G., Choquier, A., Cocirta, C., De Luca, P., Annales scientifiques, littéraires et indusrielles de l’Auvergne Dupis, A., Gay, I., Mathey, E. and Guenet, P. �������������������1995: Effets clima- XXVI, 81–229. tique et anthropique sur l’érosion mécanique en montagne cristal- Pommerol, F. 1877: Existence de l’Homme en Auvergne à l’époque du line de région tempérée pendant le Tardiglaciaire et l’Holocène. renne et des volcans à cratères. Comptes rendus de l’Association 516 The Holocene 20(4)
Française pour l’Avancement des Sciences, 5ème session, 11ème Trément, F., Argant, J., Bréhéret, J.-G., Cabanis, M., Dousteyssier, section: Anthropologie, 661–68. B., Fourmont, A., Fournier, G., Liabeuf, R., Loison, G., López- Pouenat, P. 2007: Un tesson Villeneuve-Saint-Germain dans un lot Sáez, J.-A., Macaire, J.-J., Marinval, P., Mennessier-Jouannet, épicardial. Archéopages 19, 74–75. C., Milcent, P.-Y., Prat, B., Rialland, Y. and Vernet, G. 2007b: Prat, B. 2006: Systèmes agropastoraux et milieux périurbains en Basse Un ancien lac au pied de l’oppidum de Gergovie (Puy-de- Auvergne au cours des trois derniers millénaires: contribution de Dôme): interactions sociétés-milieux dans le bassin de Sarliève à l’analyse palynologique à l’étude des interactions sociétés-milieux. l’Holocène. Gallia 64, 289–351. PhD thesis, University Blaise Pascal of Clermont II, 368 pp. Ursic, S.J. and Dendy, F.E. 1965: Sediment yield from small water- Ruddiman, W.F. 2003: The anthropogenic greenhouse era began thou- sheds under various land uses and forest covers. In Proceedings, sands of years ago. Climatic Change 61, 261–93. federal interagency sedimentation conference, 1963. US Trément, F., Argant, J., Bréhéret, J-G., Cabanis, M., Dousteyssier, B., Department of Agriculture Miscellaneous Publication 970, Fourmont, A., Fournier, G., López Sáez, J.A., Macaire, J.-J., Prat, 47–52. B. and Vernet, G. 2005: Paysages et peuplement dans le bassin de Valero-Garcés, B.L., Delgado-Huertas, A., Navas, A., Machin, J., Sarliève du Néolithique au Moyen Âge (Puy-de-Dôme, France). Gonzales-Samperiz, P. and Kelts, K. 2000: Quaternary paleo- Eléments pour un nouveau modèle socio-environnemental. In Temps hydrological evolution of a playa lake: Salada Mediana,central et espaces de l’Homme en Société, analyses et modèles spatiaux en Ebro Basin, Spain. Sedimentology 47, 1135–56. archéologie. Actes de la XXVème rencontre intern. d’archéologie van Geel, B., van der Plicht, J., Kilian, M.R., Klaver, E.R., Kouwenberg, et d’histoire d’Antibes, sous la dir. J.-F. Berger,F. Bertoncello, F. J.H.M., Renssen, H., Reynaud-Farrera, I. and Waterbolk, H.T. Braemer, G. Davtian, et M. Gazenbeek. APDCA, 485–98. 1998: The sharp rise of dC14 ca. 800 cal BC: possibles causes, Trément, F., Loison, G., Argant, J., Bréhéret, J.-G., Cabanis, M., related climatic teleconnections and the impact on human envi- Dousteyssier, B., Fourmont, A., Fournier, G., Liabeuf, R., ronments. Radiocarbon 40, 535 – 50. Lopez-Saez, J.-A., Macaire, J.-J., Milcent, P.-Y., Prat, B., Vernet, G. and Raynal, J.P. 2000: Un cadre téphrostratigraphque réac- Rialland, Y. and Vernet, G. 2006: Interactions sociétés-milieux tualisé pour la préhistoire tardiglaciaire et holocène de Limagne en Grande Limagne du Néolithique à l’époque romaine. Apport (Massif central, France). Comptes Rendus Academie des Sciences des recherches interdisciplinaires conduites dans le bassin de Paris 330, 399–405. Sarliève (Puy-de-Dôme). In Fouéré, P., Chevillot, C., Courtaud, Vernet, G., Caillat, P., Mennessier-Jouannet, C., Parent, D., P., Ferullo, O. and Leroyer, C., editors, Paysages et peuple- Rebiscoul, A., Richier, A., Seris, D., Vernet, B. and Wittmann, ments. Aspects culturels et chronologiques en France méridi- A. 2005: Rapport final d’opération de diagnostic et de fouille onale. Actualité de la recherche. Actes des VIe rencontres archéologique portant sur le bassin de Sarliève, Grande Halle méridionales de Préhistoire récente (Périgueux, 14–16 octobre d’Auvergne. INRAP, Direction interrégionale Rhône-Alpes/ 2004), coédition ADRAHP-PSO, 11–32 (Préhistoire du Sud- Auvergne, Centre archéologique régional d’Auvergne, 3 vol. Ouest, 11e supplément). Von Grafenstein, U., Erlenkeuser, H., Brauer, A., Jouzel, J. and Trément, F., Mennessier-Jouannet, C., Argant, J., Bréhéret, J.-G., Johnsen, S.J. 1999: A mid-European decadal isotope-climate Cabanis, M., Dousteyssier, B., Fourmont, A., Lopez-Saez, record from 15,500 to 5000 years BP. Science 284, 1654–57. J.-A., Macaire, J.-J., Prat, B. and Vernet, G. 2007a: Le bassin Walling, D.E. and Webb, B.W. 1986: Solutes in river systems. In de Sarliève: occupation du sol et paléo-environnement à l’Age Trudgill, S.T., editor, Solute processes. Wiley, 251–327. du Fer. In Mennessier-Jouannet, C. and Deberge, Y., editors, Wilkinson, B.H. 2005: Humans as geologic agents: a deep-time L’archéologie de l’Age du Fer en Auvergne. Actes du XXVIIe perspective. Geology 33, 161–64. colloque international de l’AFEAF (Clermont-Ferrand, 29 Zolitschka, B. 1998: A 14,000 year sediment yield record from mai-1er juin 2003). Thème régional. Lattes, Monographies western Germany based on annually laminated lake sediments. d’archéologie méditerranéenne, 385–400. Geomorphology 22, 1–17.